Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 07 Jun 2007 22:04:55 -0400
From:      Matthew Hagerty <matthew@digitalstratum.com>
To:        rick-freebsd@kiwi-computer.com
Cc:        freebsd-hardware@freebsd.org
Subject:   Re: Promise TX2300 array not detected.
Message-ID:  <4668B947.6070408@digitalstratum.com>
In-Reply-To: <20070607195828.GA368@keira.kiwi-computer.com>
References:  <4662E72B.70003@digitalstratum.com> <4662F5BF.4090709@razik.de> <4663496A.40202@digitalstratum.com> <466718DC.2030600@razik.de> <46674449.6090109@digitalstratum.com> <46678017.6080602@fluffles.net> <21250.192.85.50.1.1181233456.squirrel@mundomateo.com> <20070607164653.GB95991@keira.kiwi-computer.com> <25389.192.85.50.1.1181242998.squirrel@24.56.193.117> <20070607195828.GA368@keira.kiwi-computer.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Rick C. Petty wrote:
> On Thu, Jun 07, 2007 at 03:03:18PM -0400, Matthew Hagerty wrote:
>   
>> My reasoning for a hardware RAID is so I can set it and forget it.  If a
>> drive fails (I'm setting up a mirror), I want to be able to just swap the
>> drive and carry on without worrying about having to do something at the
>> BIOS or OS level (controller should rebuild the mirror).  Performance in
>>     
>
> I know of no RAID cards that don't require you to do something at the BIOS
> level..  on every card I've seen, if you swap a drive you must go into the
> BIOS and tell the RAID card to reconfigure the new drive.  Even on cards
> which have hot spares, if you swap the drive you need to tell it about the
> new disk.
>
> And with "real" RAID, how do you know if the RAID is having problems?  I
> haven't seen good solutions (doesn't mean they don't exist) that didn't
> require OS-level drivers and at least some OS-level interaction.  But with
> software RAID I can write my own shell scripts that detect unplugged
> drives and use S.M.A.R.T. tools to monitor drive health.
>
> Also, if you "feel" safer about somebody's random RAID card, how do you
> know there isn't a firmware bug?  At least with software RAID, the
> software is likely public and most bugs have been worked out, or you have
> some control over the operation.  Unless the RAID card comes with a
> warranty to protect your data at all costs (i.e. they will pay you the
> value of your data if anything is lost), you're taking a gamble.
>
>   
>> my case is tertiary to reliability and stability.  The TX2300 might have
>> been the wrong choice, but you would not have known from reading the
>> marketing material...
>>     
>
> Marketing also tells me that my 500 GB drive has 500 GB of storage space,
> when really it has 500 billion bytes...  All manufacturers are guilty of
> using buzzwords to sell their products.  The TX2300 *is* a RAID card after
> all, but does it require OS support?  The easy trick to distinguish between
> fakeraid and real is whether they list "Supported operating systems" or
> package a driver CD.  If it doesn't require OS support, why are they giving
> you OS drivers?
>
>   
>> The TX2300 was also blowing errors taskqueue timeout errors
>> (http://www.mail-archive.com/freebsd-hardware@freebsd.org/msg01541.html)
>>     
>
> Looks like that was in 6.1.  6.2 (-stable) at least doesn't have nearly as
> many of these issues.  In fact I haven't seen one crop up since I upgraded
> a few months ago.
>
>   
>> on the first machine I had it in, so now I'm a little skeptical about
>> using the card.
>>     
>
> Try it out and see if you have the same problems.  Perhaps it is fixed in
> 6.2.
>
> -- Rick C. Petty
>   
Rick,

Thanks for the info.  You make some good points for using a software 
RAID, I'll have to do a little more research and run some simulated 
failures so I know how to deal with the real failure when it comes.

I am installing 6.2, I was referencing that link mostly for the errors 
which are exactly what I was getting.  That probably was a little 
confusing since that link was referencing 6.1.  Sorry.  I found a thread 
that said it was not known if the problem could be solved in the 
drivers, and was due to aggressive timing on the card and that Promise 
was known for such things.  After reading that I tried the card in a 
different computer and the errors went away.  Does not leave me with a 
lot of confidence.

Matthew



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4668B947.6070408>